The technical moat for software companies is the effort to implement the logic; for instance, many companies have moats because they implement functionality that is hard to reproduce without having access to documentation needed to implement that functionality. It is also possible that there is so much functionality implemented over the years that copying it, while 'just work' is just too much work for most companies to really compete; SAP is an example.
Then there is the example of OpenAI etc; too costly to reproduce simply by not having enough compute.
Besides technical moats, you have client base moats; for instance, a trivial to implement application, like say, Resend (where you can find tons of open and paid alternatives for) did their marketing well (until they got hacked a few times that is) which gave them a moat and they still have some of it. Here is where large incubators give you some advantage; they use circlejerk to give you a lot of clients (their other companies and their clients) from the start.
Then there is a regulatory moat which is common in banking, healthcare etc; you need certificates in distinct regions etc. That makes it harder to compete without having a lot of money.
IP moats; maybe there are patents which competitors would need to work around. I'm against software patents however, they do exist and can hinder competitors to compete directly.
These are all moats which hinder competitors coming up. If a large company wants to compete with your software solution, they can copy it and go after your market. Usually they will look at your moat and just buy you out giving them the moat.
But yes, it is why VCs want you to grow as fast as possible at any cost; to 'take the market' so you have a moat in one or more of the above so in the end it's too much work to catch up and you'll be bought (exit).
Solely on technical merit, it is expected that AI will remove this moat completely over time, but we are not there yet. We don't know what software will look like at that time anyway; probably not really recognisable vs now.
does people build using javascrapt or python or java programming language have a defensible moat since everyone can use that launage and immiate your ui/ux ideas.
Your technical moat is the complexity, quality and breadth of your solution; everything can be copied but the question is the effort. It is often just simpler to just buy you out vs building; especially if it's not only complex, good quality and has a client base. See also my other reply to you; https://news.ycombinator.com/item?id=40283601
The OP question was about LLM based companies using OpenAI api's. Most (almost all) of these are thin wrappers with some prompting; most of the ones introduced here on HN or Reddit, I can outperform in minutes. I try them all to see if someone is doing something interesting, but 99% are not, they just took some SaaS starter and added AI in there and threw it on social media. Even the ones that went somewhat further are still just doing straight forward things that can be replaced trivially inside the OpenAI playground. That's no moat at all. It looks flashy but...
Not disagreeing, and it's only one aspect of Phoenix, but it might be of interest to someone reading that this LiveView-like Clojure library exists: https://github.com/tatut/ripley
most people err on the other side of the extreme by being too conservative limiting their potential based on the wrong belief that all successes are due to luck and they can't do anything about it.
Most people woule benefit by being more optimistic and taking more risks. In the modern world, most risks won't destroy you. But you will live a very mediocre life if you believe you dont have much control over external outcome or your own success
you need to qualify that with “that is actually available”. :-). A100s I hear are harder to get in bulk. But I have modest needs!
I have been using modal and vast. Vast is cheaper. Modal has some a free inclusions of $30 but that is probably $8 to get the same power in Vast. Modal resell AWS/GCP at the moment. GCP direct seems cheap enough. As does Lambda labs.
With vast some machines don’t start, so you just need to bin them and try another. For learning and non private this is acceptable. For serious stuff I think Vast lets you filter for data-centre GPUs. Modal tends to just work and lets you store the model for later more easily.
Overall: just go with vast. You boot it up and run SSH. It is a familiar experience. Very little time needed on RTFM stuff!
how can you prevent yourself from getting sued into oblivion in the future if your tech gets good? 10 min of google search was enough to find your true legal identity so it is not like you can hide.
